Research Of Spectrum Sensing In Cognitive Radio Network

The Research are supported by the National Natural Science Fundation "Research on spectrum detection based on spectrum utilization statistic" (NO. 60772110), Huawei Technology Fundation "Research about cognitive radio in MAC Layer" and Field Fundation of Beijing University of Posts and Telecommunications.As a revolutionary smart spectrum sharing technology, CR can significantly improve the spectrum utilization and receives more and more interest within these years. In the CRN, CR users can automatically sense the wireless environment, and then access the CRN within the spectrum which had been allocated to the licence users but not occupied right now. The key point in CR technology is spectrum sensing ("spectrum detection" is also used), i.e. how to find the spectrum hold and identify the licence users.In this paper, the subject background is described in the fist section, and the significance of CR is showed there. Then the spectrum sensing methods in CRN is discussed by theory and simulation following the description of weak signal detection. Next, a new spectrum detection mechanism proposed in this paper will be discussed.In section 2, Brief instruction of weak signal detection is given. Several typical detection methods are described and comparied primarily.In section 3, some local spectrum sensing schemes in cognitive radio network were introduced at first, the impact of spectrum sensing under shadowing/fading environments and noise uncertainty, as well as adaptive MAC-layer sensing algorithm was analyzed with emphasis. And the effect of the selection of data fusion and decision algorithms, shadowing correlation, the presence of untrusted users on detection performance of cooperative spectrum sensing was discussed.In section 4, a new spectrum sensing mechanism based on flexible sensing period is proposed to control the cost of the sensing in the MAC layer of cognitive radio networks. The CR users can make a balance between the unexpected time and the sensing count by utilizing the statistic information of licence users and a "period control factor", that will reduce the cost of sensing. Simulation results shows that FSP can give a better performance than the typical fixed sensing period mechanism.